Difference between revisions of "Talk:Simulating new functions"

The math looks fine to me. You can increase the accuracy of the arctan function by adding t* terms. See http://en.wikipedia.org/wiki/Taylor_series#List_of_Taylor_series_of_some_common_functions for more info. The accuracy of the square root function you once had could be increased by adding more the same lines, though I think it's accurate enough.  Of course, there's a balance between speed and accuracy, and I think it would be okay if the arctan func is accurate to 1 degree. BTW you should add that square root function back - even if it isn't used for getAngle, it's still be a very useful function. --[[User:Maian|Maian]] 19:40, 8 May 2006 (EDT)

The math looks fine to me. You can increase the accuracy of the arctan function by adding t* terms. See http://en.wikipedia.org/wiki/Taylor_series#List_of_Taylor_series_of_some_common_functions for more info. The accuracy of the square root function you once had could be increased by adding more the same lines, though I think it's accurate enough.  Of course, there's a balance between speed and accuracy, and I think it would be okay if the arctan func is accurate to 1 degree. BTW you should add that square root function back - even if it isn't used for getAngle, it's still be a very useful function. --[[User:Maian|Maian]] 19:40, 8 May 2006 (EDT)